Feature bloat: Psychology boffins find people tend to add elements to solve a problem rather than take things away Scientists working on the psychology of problem solving may have hit upon why things always seem to get more complicated. A newly uncovered heuristic – a mental shortcut or rule of thumb – shows bias towards adding features to find a solution, rather than subtracting existing features. A simple experiment in Lego has provided …
By saying the cost of adding bricks they are insinuating a list of allowable moves. Therefore in the Lego example the implication is that everything should stay where it is and you should only stabilise the roof, not redesign the building.
However, regardless of their crap experiment design I probably do still believe their hypothesis.
Yeah, I think I would also have assumed that the roof was raised up on that brick for a reason, and so also added more supports. But three total should have been enough, it wouldn't have needed four, ... or maybe a minimal solution would be just two staggered slightly given they're 2x2 not 1x1? Hmm...
I think there is still some duplo lurking in a box somewhere. Hang on whilst I go get it, and design a proper experimental procedure for stability testing.... :-)
"the roof was raised up on that brick for a reason"
Ah yes, "the reason". I get this all the time. Most recently when I was pointing out how stupid an implementation was, I got, "well I would hope they had a good reason for doing it that way". I lost my temper, big time. I still have a job though, because when it got escalated it turned out "they" did not in fact have a good reason.
Quite aside from any reason for the building design, I wonder if people are conditioned to expect missing pieces, like with a jigsaw puzzle. How much time do people spend trying to find a missing part or fix a broken part? So you see a roof with one support, it's a known pattern of "missing pieces".
That's actually a dangerous heuristic you're implying. You certainly never spent much time on a farm, for instance.
Just because YOU don't know why something was done does not mean that there was no reason. There almost certainly was. The correct question is "is the reason currently valid?"
"Just because YOU don't know why something was done does not mean that there was no reason."
I agree 100%. But you misunderstood the scenario. As the domain expert I was telling them the implementation was wrong. Which it *obviously* was, and which was agreed after escalation. Shouldn't have needed escalation, and shouldn't have passed review in the first place. But the initial response was not an actual stated reason, but a vague "some reason", which turned out not to exist.
"You don't have to walk a mile in someone else's shoes to know their laces are untied." -- long ago letter to Dear Abby advice column
By the way, I have spent time on a farm. Enough time to know there's a lot going on I don't know about. Are you saying farmer Brown can't call farmer Smith an idiot when the latter does something boneheaded for "reasons"?
Farmer Brown's free speech rights are fine so far as I am concerned.
And I also know that there were MANY times when one farmer would call another a fool or some such because they disagreed with what the other was doing. Doesn't mean they were correct, though.
But what I was talking about was the "leave open gates open & closed gates closed" rule. You don't get to change the status of the gate unless you know exactly why it was the way it was & how that reason no longer applies.
Your lack of imagination is no excuse to condemn someone else's behavior.
Certainly, if you are the SME, and someone lacking expertise attempts something that appears stupid, you are justified in challenging what's going on. But until you do? ...
I agree that all of the experimental designs the used possess implicit preconditions. A roof is designed to be at a particular height, therefore the maintenance of that feature could be assumed to be an implicit design condition. White and green squares... white is often considered to be a background colour, so is one adding white or removing green? There is an implicit precondition required that white is seen as a background colour instead of anything else like the toggle of green. And as for the numbers distraction task... that's just a fudge for an improperly designed experiment.
However, like the previous comment, I too believe them! If you have for example code that doesn't react well to a specific input condition, the all too common response is to trap that input condition rather than rethink the code so that it works correctly with a wider range of inputs. Thus we end up for example with UIs that utterly refuse to allow one to progress to the next stage without being fully completed even when certain fields may not have a value, so some users "make up rubbish" as the value for a stubborn field whereas a NULL would be perfectly OK if only someone had allowed that in the backend to begin with.
This is like the old days of trying to write code that reads the keyboard as the user types by sampling the serial input... easy to do until the user presses the arrow keys to move the cursor around the text and the characters come faster than the routine is sampling and start getting dropped - what's the solution?
The standard solution was to increase the serial character sampling rate ... but the better solution was to move to an interrupt service routine - that always worked 100% whereas boosting the sample rate meant the CPU spent too much time waiting to a character as the user typed and dropped them occasionally.
It is about implicit requirements. Were the participants given a functional specification ? Would it have been OK to remove all of the bricks, move the man and put the roof on the floor ?
As any barrister knows you greatly influence the answer that you get by the way that you ask the question.
This post has been deleted by its author
How many software packages had all the useful features added between version 1 and (3 to 5), and subsequently later versions kept adding functionality that REDUCED the overall performance / usefulness of the software?
Sir! Sir! I know this sir! That's a trick question sir! All of them sir!
No software has ever been improved after version 5.
I remember a saying about writing software. There's a theory that any significant block of code contains at least one error. There's also a theory that any significant code can be simplified by removing one line. Therefore, by reduction, every single computer program can be reduced to a single line of erroneous code.
"Defaulting to searches for additive changes may be one reason that people struggle to mitigate [...] damaging effects on the planet"
So maybe less consumption, travelling, fuel burning, waste production etc., rather than allowing consumption, travelling, fuel burning, waste production etc. to increase for ever but doing it "more greenly", might be a good idea?
Councils are keen on doing what councils do, to justify their existence.
I was living in London when there was the vote as to whether there should be a mayor. I voted against - but I was on the losing side.
Since then, millions have been spent on new office buildings and thousands of staff. Mostly what they do is to introduce new charges (e.g. congestion charge) to offset these costs.
"What this country needs is more politicians!" -- said nobody ever.
In the upcoming local election set there's one for a mayor of West Yorks. Apparently there are a substantial list of candidates but i doubt there'll be an option for "No, I don't want one.". An even more stupid idea than "Krklees"; we probably have more in common with rural parts of Derbyshire than we do with Leeds, Bradford etc.
To be fair that is being driven by attempts to rationalise in order to save money. Inevitably it concentrates the complexity in the hope that this will reduce the number of skilled workers required: JIT for the informational age. And indeed some of the ideas are reasonable: encapsulation, repeatability and reliability. But the motivation behind some of the decisions is certainly questionable.
"As with many heuristics, it is possible that defaulting to a search for additive ideas often serves its users well,"
Yeh, but the model is big already, it is built on the work of a *lot* of people, lots of bricks have been placed, so who are you, *one* person, to dare question if any or all those people screwed up and misplaced their bricks! I think that's the logic there. Adding 3 blocks adds to their work, removing 3 blocks detracts from their work and risks criticism. Hence additive is easier.
Look at a more concrete example, retrograde motion of planets:
https://earthsky.org/space/what-is-retrograde-motion
A bunch of very clever mathematicians built a model to predict the apparent backwards loops-da-loop the planets took, given the earth is the center of the universe. It must be true because observation fits the model and the earth must be the center of the universe. Some fixes were needed, so crap was piled on crap, and it became the science of the day.
Even if it a ridiculous model with planets doing dances. "Who am I to question the complex, detailed and working model of all these smart people?" thinks your astronomer of the day. It's amazing how such a ridiculous system could stand as true for so long, yet it did.
Then Copernicus and later Kepler *did* question that model, and did point out the *sun* is at the center and the planets move around it, and the loop-da-loop of the planet became a simple net motion of the observed planet and the observer planet earth. Suddenly the stupidity of the original model is apparent, at least to *some* people, but not *all* people.
I think there are 3 groups in any argument: The proponents of a new model, the opponents who defend the existing model, and the observers who sit on the sidelines and watch.
Arguments aren't really about winning over the opponent and getting them to change their mind. They simply don't change their mind. They double down. Hurl insults and crap and misdirection. The more their self worth is tied to the broken model, the nastier they get defending it. Self preservation is an understandable human trait.
No, the purpose of an argument is to plant the flaw in the *observer* and flip the balance of understanding away from the broken existing model. If you can make it so clear and simple that the undeniable nature of the flaw is apparent to a lot of people, you can flip the understanding quicker, even as the opponents seek to obfuscate the mistake.
The balance of understanding shifts as new scientists come in, realize they can make their name and a flip in understanding happens, and science moves off in a new direction.
"Arguments aren't really about winning over the opponent and getting them to change their mind. "
The way I've heard it described is that most people don't change their minds, but old people with old ideas die off and the younger generations learn and accept the newer models until over the course of a couple of generations there are no more defenderds of the old model.
It has a lot to do with ego and people being personally attached to their past work
think there are 3 groups in any argument: The proponents of a new model, the opponents who defend the existing model, and the observers who sit on the sidelines and watch..
At risk of adding..
There's 4 groups. There's also the ones that think "Hmm maybe they have a point, let's have a look". but they get shouted down by the oppenents for not opposing. And the proponents for doubting.
Writing a paper asserting people are hardwired to create more complexity, while working in a field in which your worth is measure by the number of papers one writes.
Back to my day job of creating more product...
We are all hamsters on the wheel, my friends.
This post has been deleted by its author
Yes, the design pre-implies that the the roof has to stay where it is. The task presented was: stablise this roof four blocks above ground level. If I contracted a builder to repair my roof on my two-storey house and he implemented it by converting my house to a one-storey house I would be fairly annoyed.
Hopefully your house isn't made of Lego, and situated in a lab where any idiot might be encouraged to remodel it. And equally hopefully the builder has very detailed instructions from you, the architect and the engineer before he starts the renovation. If not, I submit your new bungalow is clearly what you asked for.
My happiest days, picked up a huge mountain of Solaris code, it crashed, it didn't work, it was impossible to understand or debug. After 2 weeks i deleted it all and spent a happy 8 hours writing a new solution that worked first time without crashing, it was only a couple of hundred lines. Recently did similar with some asn1 decoding code, table driven and less lines than the original had classes
Works with people, too.
I was brought in to down-size a company once. Took three months, but I finally got rid of 20% of the workforce, all in middle management (and nobody else). Company dropped from about 1000 people to about 800, with a good rise in productivity. Profits rose by almost triple the prior year's cost of keeping the middle management employed.
Easy way to start implementing this at your corporation: Get rid of any middle management that does nothing but fiddle about with Power Point and attend meetings. Then get rid of any middle management who complains about those cuts. That should get you into the 15% range or so, you're on your own for the rest.
This post has been deleted by its author
At my “my” TBTF, that would be all of them, all of the 3 layers. There’s even that one guy whose weekly progress report runs to 40 slides. Which is probably what he did all week (Which is probably also for the benefit of the organisation).
Those meetings are what kept me from pursuing a career in management, 10 years of PowerPoint tenure before getting getting the dread powers of cleansing and purifying :).
There’s a re-org coming, September, the rumours say. The spicy part it the old boss left in a huff & puff in the middle of his contract. A new boss is being acquired, interim boss is doing not much (that I can see), the 3-months “window of change” has almost passed. Now, of course the rumours were wrong about the last 3 bosses.
The next one might also just add another layer of PowerPoint perceptrons in order to make the organisation function better. Seems to be the default setting for executive level Leadership these days.
Exhibit A, Risk models used by a major power utility. The Regulators and stakeholders criticises it for being very difficult to understand as an outsider. Dieter Helm's cost of Energy review criticised the whole framework for much the same reasons.
So, what do we do? Make it MORE complicated; publishing most of the methodology in the name of "transparency"; yet ever more unclear to even the people who live with having to use and explain it.
Can you see where this is going next?
"the first question to ask is whether the roof has to be at that height"
I used to have Legos as a child, I think some here have forgotten the point. The first question I would ask is, who is paying the 10 cents per part? And the second question is, how many parts are available? Because of course the ideal solution would use every available brick, including any not being used by the other players.
Look at tax codes for example. There may have been a reason for every single credit or deduction to be added to the code at one time (even if the reason was "because the right people were bribed") but when it doesn't work due to i.e. allowing highly profitable companies or very wealthy individuals to pay little or no tax, the response is "we need to add to the code to address this situation". Rather than look what credits/deductions allow them to do this and ask "is this still necessary" or "can its scope be narrowed"?
So people are talking about revenue taxes on business and wealth taxes on billionaires, rather than addressing the credits/deductions that allow them to do this in the first place.
Which is why I've always thought that redoing the tax code should take a clean sheet approach where a high bar to justify including any deductions or credits (beyond a phase out period to address inequities due to changing things people may have planned years for in advance like mortgage interest deductions in the US)
You see it even more in the legislature.
Laws are rarely repealed. Instead they are amended or a new law is passed which adds to rather than replaces the old law.
It is left to the bureaucracy to figure out how to implement the laws and to the courts to decide whether they got it right.
William of Ockham's Razor - "Numquam ponenda est pluralitas sine necessitate" ("Plurality must never be posited without necessity")
Take the three blocks from the centre and use them to stabilise the three corners, leaving a hollow tower.
OR
Take the four blocks from the middle of the upper rows, put one in the centre and stabilise the corners with the other three.
Major flaw in constructing the test, It assume the person does not assume the building is designed to be a specific height with a overhang etc. the simplest fix would be to remove the roof in its entirety. The brick may fall on the head of the innocent lego person but that was not a condition of the fix.
The theory may be correct, but the proof proposed by this test here is flawed on many levels.
A long time ago (1975?), I worked on the "Europeanisation" of the Xerox 800 (word processor based on the 8080).<P>
In those days, it was quite hard to get American semiconductors in Europe, so we had to replace the design at PCB level
with a locally designed equivalent.<p>
In most cases, the European design used fewer parts, which we attributed to solving problems by removing
parts while Americans solved them by adding parts.
It's called "There was an old lady who swallowed a fly"
One of the biggest issues when faced with a problem and a solution which doesn't work is the tendency to layer on more+more "fixes" rather than stepping back and seeing if there's a better way of approaching the original problem. People get tunnel vision about things
Different people have different rules about what is OK to do to another’s creation - and the rules are largely subconscious. They try to avoid that bias: “All the participants were told they could alter the structure however they wanted to.”
But that isn’t quite the same as giving “ownership” of the tower. It is hard to guess what one would do as a participant, but I suspect I would feel it wasn’t “my” tower, so my scope of changes would be more limited.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
